
import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
import seaborn as sns
sns.set_style('darkgrid')

import plotly
import plotly.express as px

%matplotlib notebook


churn_data = pd.read_csv('../data/churn.csv')
churn_data.head()


plt.figure(figsize=(6,5))
sns.countplot(x='Exited',  data=churn_data[['Exited']])
plt.title(label='Cоотношение ушедших и лояльных клиентов')
plt.show()


df = churn_data[churn_data.Balance > 2500]

fig, axes = plt.subplots(figsize=(7, 5))
sns.histplot(
    data=df,
    x='Balance',
    bins=25,
    kde=False,
    ax=axes
);
axes.set_title('Распределение баланса пользователей с балансом > 2500 usd', fontsize=10)
plt.show()
print(df.Balance.describe())

count      6383.000000
mean     119827.493793
std       30095.056462
min        3768.690000
25%      100181.975000
50%      119839.690000
75%      139512.290000
max      250898.090000
Name: Balance, dtype: float64


fig = plt.figure(figsize=(7, 5))
boxplot = sns.boxplot(data=churn_data, x='Exited', y='Balance')
boxplot.set_title('Boxplot', fontsize=10)
boxplot.set_xlabel('Тип клиента')
boxplot.set_ylabel('Balance');


fig = plt.figure(figsize=(7, 6))
boxplot = sns.boxplot(data=churn_data, x='Exited', y='Age')
boxplot.set_title('Boxplot', fontsize=10)
boxplot.set_xlabel('Тип клиента')
boxplot.set_ylabel('Возраст клиента');


fig = plt.figure(figsize=(7, 6))
sns.scatterplot(data=churn_data, x="EstimatedSalary", y="CreditScore", hue='Exited')
fig.show()


fig = px.pie(churn_data[churn_data['Exited']==1], 
             values='Exited', names='Gender', 
             title='Left the bank',
             height=500,
             width=700
            )
fig.show()


def map_numbofprod(numbofprod):
    if numbofprod == 1:
        return "NumberofProd-1"
    elif numbofprod == 2:
        return "NumberofProd-2"
    elif numbofprod == 3:
        return "NumberofProd-3"
    elif numbofprod == 4:
        return "NumberofProd-4"
    
numofproducts = [len(churn_data[churn_data["NumOfProducts"]==n]) \
                 for n in np.sort(pd.unique(churn_data["NumOfProducts"]))
                ]
numofproducts = pd.DataFrame(np.reshape(numofproducts, (len(numofproducts),1)), columns=['Total'])

churn_data["NumOfProducts"] = churn_data["NumOfProducts"].apply(lambda x: map_numbofprod(x))
bar_data = churn_data.groupby(
    by='NumOfProducts',
    as_index=False
)[["Exited"]].sum()

bar_data = pd.concat([bar_data, numofproducts], axis=1)
bar_data = bar_data.rename(columns={"Exited": "Exited-1"})
bar_data['Exited-0'] = bar_data.Total - bar_data['Exited-1']

fig = px.bar(bar_data, x="NumOfProducts", y=["Exited-1", "Exited-0"], title="Number of Prod and Exited",
            height=500,
             width=700)
fig.show()


def map_isactivemember(isactivemember):
    if isactivemember == 1:
        return "ActiveMember"
    elif isactivemember == 0:
        return "Not ActiveMember"
    

IsActiveMember = [len(churn_data[churn_data["IsActiveMember"]==n]) \
                 for n in np.sort(pd.unique(churn_data["IsActiveMember"]))
                ]
IsActiveMember = pd.DataFrame(np.reshape(IsActiveMember, (len(IsActiveMember),1)), columns=['Total'])

churn_data["IsActiveMember"] = churn_data["IsActiveMember"].apply(lambda x: map_isactivemember(x))
bar_data = churn_data.groupby(
    by='IsActiveMember',
    as_index=False
)[["Exited"]].sum()

bar_data = pd.concat([bar_data, IsActiveMember], axis=1)
bar_data = bar_data.rename(columns={"Exited": "Exited-1"})
bar_data['Exited-0'] = bar_data.Total - bar_data['Exited-1']


fig = px.bar(bar_data, x="IsActiveMember", y=["Exited-1", "Exited-0"], 
             title="Active Client and Exited", 
             height=500,
             width=700
            )
fig.show()


choropleth_data = churn_data.groupby(
    by='Geography',
    as_index=False
)[["Exited"]].sum()

fig = px.choropleth(
    data_frame=choropleth_data, 
    locations="Geography", 
    locationmode = "country names",
    color="Exited", 
    range_color=[0, 800],
    title='Exited clients',
    width=700, 
    height=500, 
    color_continuous_scale='Reds' 
)
fig.show()


def get_credit_score_cat(credit_score):
    if credit_score >= 300 and credit_score < 500:
        return "Very_Poor"
    elif credit_score >= 500 and credit_score < 601:
        return "Poor"
    elif credit_score >= 601 and credit_score < 661:
        return "Fair"
    elif credit_score >= 661 and credit_score < 781:
        return "Good"
    elif credit_score >= 781 and credit_score < 851:
        return "Excellent"
    elif credit_score >= 851:
        return "Top"
    elif credit_score < 300:
        return "Deep"    
    

churn_data['CreditScoreCat'] = churn_data['CreditScore'].apply(lambda x: get_credit_score_cat(x))
 
    
pivot_table = churn_data.pivot_table(
    values='Exited',
    index='CreditScoreCat',
    columns='Tenure',
    fill_value=0
)


fig = plt.figure(figsize=(6, 5))
heatmap = sns.heatmap(data=pivot_table, annot=True, cmap='YlGnBu')
heatmap.set_title('Categories of Exited clients', fontsize=10);
fig.show()

	RowNumber	CustomerId	Surname	CreditScore	Geography	Gender	Age	Tenure	Balance	NumOfProducts	HasCrCard	IsActiveMember	EstimatedSalary	Exited
0	1	15634602	Hargrave	619	France	Female	42	2	0.00	1	1	1	101348.88	1
1	2	15647311	Hill	608	Spain	Female	41	1	83807.86	1	0	1	112542.58	0
2	3	15619304	Onio	502	France	Female	42	8	159660.80	3	1	0	113931.57	1
3	4	15701354	Boni	699	France	Female	39	1	0.00	2	0	0	93826.63	0
4	5	15737888	Mitchell	850	Spain	Female	43	2	125510.82	1	1	1	79084.10	0

¶

¶

¶

¶

¶

¶

¶

¶

¶